Skip to main content Accessibility help


  • Access
  • Cited by 5



      • Send article to Kindle

        To send this article to your Kindle, first ensure is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle. Find out more about sending to your Kindle.

        Note you can select to send to either the or variations. ‘’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

        Find out more about the Kindle Personal Document Service.

        Microchemical Study of Pigments and Binders in Polychrome Relics from Maiji Mountain Grottoes in Northwestern China
        Available formats

        Send article to Dropbox

        To send this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Dropbox.

        Microchemical Study of Pigments and Binders in Polychrome Relics from Maiji Mountain Grottoes in Northwestern China
        Available formats

        Send article to Google Drive

        To send this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Google Drive.

        Microchemical Study of Pigments and Binders in Polychrome Relics from Maiji Mountain Grottoes in Northwestern China
        Available formats
Export citation


In this study, an integrated analytical method was developed to investigate the composition of both the inorganic pigments and organic binders of polychrome relics in Maiji Mountain Grottoes in northwestern China. Cross-sections of each sample were prepared at the beginning of the study, and all experiments were carried out on these cross-sections. Polychromic structures were revealed by optical microscopy and scanning electron microscopy-backscattered electron imaging. Inorganic materials were determined by using SEM coupled with an energy dispersive spectrometer and μ-Raman spectrometer, whereas organic materials were identified by staining techniques and highly sensitive and specific immunofluorescence microscopy. Data showed that the red colors are attributed to one or two pigments of red ochre, cinnabar, and minium; the blue pigment is natural lazurite; the green pigment is ascribed to atacamite; the white color is attributed to potassium feldspar; and the black surface is formed by the discoloration of minium to plattnerite under the influence of environmental factors. Regarding organic binders used in painting and preparation layers, mammalian animal glue and chicken egg white were both found alone or in mixture. Finally, the conclusion is made that the Secco technique is employed in polychrome relics from Maiji Mountain Grottoes.


There is a great variety of exquisite polychrome relics in China. In order to reveal original materials used in ancient times, and to disclose ancient painting techniques, exhaustive investigation of chemical composition of both inorganic and organic compounds in polychrome relics is necessary. The generally complex nature of these materials, and small or even unavailable samples because of their preciousness make the task challenging. Fortunately, microchemical analytical techniques, including both destructive and nondestructive analytical methods such as polarized light microscopy (Piovesan et al., 2011), X-ray fluorescence (XRF) (Rosi et al., 2009), energy dispersive X-ray detector (Mazzocchin et al., 2003), X-ray diffraction (XRD) (Franquelo et al., 2012), Raman spectroscopy (RS) (Burgio & Clark, 2001), Fourier transform infrared spectroscopy (FTIR) (Cappitelli & Koussiaki, 2006; Wang et al., 2014) and gas chromatography (GC) (Cappitelli & Koussiaki, 2006), have been developed rapidly during the last century and make the task easier.

Polychrome relics are often composed of several layers, including the mud ground layer, white priming layer, original paint layer, and sometimes additional repaint layers. In order to get more information regarding the composition and morphology, micro-destruction for cross-section analysis is unavoidable though in situ techniques, such as portable XRF (Bardelli et al., 2011) and diffuse reflectance FTIR (Navas et al., 2008), promise to identify compounds on the surface of artworks (Mancini et al., 2012; Veneranda et al., 2014). Cross-sectional observation by optical microscopy (OM) and scanning electron microscopy coupled with backscattered electron (SEM-BSE) imaging are effective methods to examine polychromic structures. In the analysis of inorganic materials, complementary analyses with SEM coupled with an energy dispersive spectrometer (SEM-EDS) and micro-Raman spectroscopy (μ-RS) are ideal because they provide information about elements and compounds, respectively. Pigments in Maiji Mountain Grottoes have been intensively investigated by XRD (Zhou, 1991) and these microchemical techniques provide some detailed information.

Binders used in ancient Chinese polychrome relics are normally natural organic materials. Wang et al. (2014) used FTIR to investigate the gilding sculptures from a temple and found that drying oil, Chinese lacquer or animal glue were used in the adhesive layer to joint gold leaf with the preparation layer. Wei et al. (2012) used pyrolysis gas chromatography mass spectrometry (py-GC-MS) and GC-MS to analyze Western Han’s (206 BC–8 AD) polychromic terracotta army and found that animal glue was applied as a binder. Bonaduce et al. (2008) and Hu et al. (2015) found the binder used in Qin Shihuang’s Terracotta Army was egg.

In this work, immunofluorescence microscopy (IFM) was used to detect binders in polychrome relics. IFM, which uses a highly specific and sensitive immunological antigen–antibody reaction to detect the presence of an antigen, is a fast and cost-effective technique that can recognize and locate antigens accurately (Vagnini et al., 2008). However, due to nonspecific fluorescence, the application of IFM in conservation science is scarce. To suppress the light scattering phenomenon from the sample surface, Cartechini et al. (2010) suggested several methods such as the use of confocal fluorescence microscopy in the collection of fluorescence from given depth and long lifetime fluorescent labels coupled with time-gated analysis. However, due to the high cost of the apparatus, it is not practical for common museums to do routine analysis with the above-mentioned methods. In this work, the simple and cost-effective methods, Oil Red O and Sypro Ruby stain, were applied with IFM to determine binders in polychromic samples. These two stain tests made general classification of oily and proteinaceous materials possible (Sandu et al., 2012 a ; Kuckova et al., 2013) and the consistency of the analytical results of Sypro Ruby staining and IFM increased the accuracy of authenticity studies.

Maiji Mountain Grottoes is located 45 km southeast of Tianshui city, in southern Gansu province. “Maiji” means wheat-rick in Chinese. As the mountain where the grottoes are located looks like stacks of wheat-rick, it was called Maiji Mountain by the natives. Construction of the Maiji Mountain Grottoes began in the early fifth century and was rebuilt and renovated during the next 1,600 years (Dong, 1983; Huang, 1989). Nowadays, there are 221 grottoes and 3,988 statues and murals that cover >979 m2 (Wei, 2011). Owing to such a large number of elegant cultural relics, Maiji Mountain Grottoes is honored as the “Oriental Sculpture Gallery”. In the 38th session of the World Heritage Committee held on June 22, 2014, the application for adding part of “Silk Road” to the World Cultural Heritage list was jointly submitted by China, Kazakhstan, and Kyrgyzstan, and approved by the committee. Maiji Mountain Grottoes was included as one of the important components in this application, indicating that its significance was approbated by the international community. Consequently, the study of polychrome relics from Maiji Mountain Grottoes is meaningful for deepening the understanding of ancient techniques and inheriting these cultural relics.

In this work, ten fragments from Maiji Mountain Grottoes were fully examined by a series of techniques. The polychromic structures were studied by cross-sectional examination including OM and SEM-BSE observations and inorganic materials are identified by SEM-EDS and Raman analyses. Then traditional staining techniques and IFM were used in sequence to disclose proteinaceous binders. All these techniques were examined on the cross-sections to reveal the distribution of inorganic and organic materials. The research work in the present study was carried out in order to determine the materials employed for the elaboration of the original work, and to deeply understand the techniques used in Chinese traditional painting. Considering that Maiji Mountain Grottoes is a typical heritage on the Silk Road, the results of this study on polychrome relics are representative of classical grotto arts in ancient northwest China of that period.

Materials and Methods


A very small amount of painted materials (each the size of a few mm2, Table 1) were taken from damaged areas or debris, which had already fallen to the ground from different caves of the Maiji Mountain Grottoes. Each of the samples included both the surface pigment layer and ground layer. The descriptions of samples are listed in Table 1.

Table 1 Description of the Samples from the Maiji Mountain Grottoes.

Cross-Section Preparation

A small part of the samples was carefully separated and embedded in acrylic resin to prepare the cross-sections. Heat or ultraviolet (UV) was not used during solidification since they could denature the proteinaceous binders in the samples. After curing, the resin block was cut into small pieces, which was then glued on glass slides and polished with a series of increasingly finer grades of micromesh (600, 1,000, 2,000, and 5,000 mesh) to make the surface smooth and the samples exposed.

Analysis Methods


Cross-section images under visible light were observed with a Keyence VHX-700FC digital microscope (Osaka, Japan) using a VH-Z100R objective allowing magnification from 100×–1,000×. Fluorescence in cross-sections was characterized by a Nikon Eclipse Ti inverted microscope (Tokyo, Japan) provided with a C-SHG 1 mercury lamp (Nikon, Tokyo, Japan). Filter B-2A (EX 450–490, DW 505, BA 420) was used for observing the fluorescence to generate pure blue light. The pictures were recorded with a LH-M100CB-1 digital camera (Nikon).

SEM-BSE Observation and SEM-EDS Analysis

These analyses were carried out using a Hitachi SU8010 SEM equipped with an IXRF Systems energy dispersive X-ray spectrometer (EDS; IXRF Systems, Austin, America). SEM-BSE was used for taking micrographs of cross-sectional paintings. EDS analysis was carried out in low vacuum (water vapor pressure in the chamber set at 100 Pa) and an accelerating voltage of 15 kV.

μ-RS Analysis

Horiba Scientific XploRA Raman spectrometer (Horiba, Kyoto, Japan) configured with an Olympus BX41 microscope was employed for the μ-RS analysis with 10× and 100× objectives. Two excitations with wavelengths of 532 and 785 nm were used. The maximum output powers were 20–25 mW (532 nm) and 90–100 mW (785 nm), respectively. The instrument was calibrated daily with the 520 cm−1 silicon Raman band. The Stokes Raman spectra were obtained over the spectral range of 100–2,000 cm−1 and the resolution was kept at 3.2 cm−1. Finally, the spectra were compared with the Horiba commercial spectral database and databases in the literature.

Traditional Staining

Proteinaceous and oily materials were identified with fluorescent staining with Sypro Ruby and visible staining by Oil Red O, respectively. Sypro Ruby, obtained from Molecular Probes (Eugene, OR, USA), is a biomedical non-covalent stain generally used in proteomics for fluorescence mapping of proteins. It is a ready-to-use solution and has good sensitivity and selectivity down to nanogram levels (Sandu et al., 2012 a ). The analytical procedure comprised the following steps: the cross-section was stained with a drop of Sypro Ruby for 30 min before rinsing with distilled water and then observed immediately with a fluorescence microscope and blue light.

Oil Red O was purchased from Sigma-Aldrich (St. Louis, MO, USA) and prepared according to an already tested recipe: 0.125 g of agent was mixed with 25 mL isopropyl alcohol, filtered after 1 h, and diluted with distilled water in 3/2 (v/v). One drop of this mixture was placed on the cross-section and incubated for 10 min. Then the cross-section was washed with isopropyl alcohol/water 3/2 (v/v) and distilled water in sequence to remove unbound dye compounds and observed by OM (Sandu et al., 2012 b ).

IFM Analysis

Animal glue and egg white, which were most likely used as proteinaceous binders in ancient Chinese relics were identified by the existence of their corresponding main components, namely mammalian collagen and chicken ovalbumin, respectively. Anti-collagen antibody (rabbit polyclonal antibody to collagen type I, AB749P) was received from Millipore (Temecula, CA, USA) and anti-ovalbumin antibody (mouse monoclonal antibody to ovalbumin, A6075) was purchased from Sigma-Aldrich (St. Louis, MO, USA). The secondary antibodies goat anti-rabbit and anti-mouse IgG (H+L) both conjugated with Alexa Fluor 488 were received from Invitrogen-MP (Carlsbad, CA, USA). Phosphate-buffered saline solution (PBS, 150 mM NaCl, 5.2 mM Na2HPO4, 1.7 mM KH2PO4, pH 7.4, 0.2% Tween20) was used to dilute antibodies and for washing steps. The blocking solution (P0102) used in IFM was purchased from Beyotime (Beijing, China). The procedure has been reported in our previous work (Hu et al., 2015). Briefly, the cross-sections were incubated in the blocking solution for 1 h at room temperature to avoid nonspecific reactions with the sample. The primary antibody diluted with PBS at 1/200 (v/v) was added and incubated overnight at 4 °C. Then the fluorescein-labeled secondary antibody in the dilution of 1/500 (v/v) was added and incubated in the dark at room temperature for 2 h. The samples were observed with fluorescence microscope and blue light. Between each step, the cross-sections were rinsed with 5 mL PBS for three times.

Analytical Sequence

In order to identify the structure of each sample and the composition of inorganic materials and organic binders, the above-mentioned analytical methods were carried out with the following strategy. First, the sample was observed by microscopy and a part of the sample was embedded in acrylic resin to prepare the cross-sections. Then, the cross-sections were observed by OM and SEM-BSE to identify the color and structure of each layer. The inorganic pigment was first tested by SEM-EDS to determine element composition and then analyzed by μ-RS to identify the compounds. As proteinaceous and oily materials were most likely used as binders in ancient Chinese polychrome relics, the staining of Sypro Ruby and Oil Red O was carried out in sequence on cross-sections to get the distribution of these two materials and further IFM was performed to get the precise information of proteinaceous binders.

Results and Discussions


The cross-sectional examination by OM and SEM-BSE micrographs for samples with different colors is shown in Figures 1, 3, and 5. The layer number of each sample is given in the OM micrograph. The measurement spots where the elemental content (in wt%) is determined by SEM-EDS analysis (see Supporting 1 in Supplementary Material) have been marked with round dots or arrows in the SEM-BES micrographs. Identification of a certain pigment is based on elemental composition from EDS analysis and characteristic peaks from μ-RS (Figs. 2, 4). The summary of color stratigraphy, polychromic structures, and the pigments is shown in Table 2.

Supplementary Supporting 1

Supplementary Supporting 1 can be found online. Please visit

Figure 1 The cross-sections of red samples (a) MJ3, (b) MJ4, (c) MJ6, and (d) MJ10 by optical observation (left) and BSE observation (right), the spots where wt% are given are marked with bullets in the BSE images.

Figure 2 Raman spectra of (a) red ochre in layer 2 of MJ3, (b) lead white and red ochre in layers 2 and 4 of MJ4, (c) cinnabar in layers 7 and 9 of MJ6, (d) minium in layers 3 of MJ10 and (e) plattnerite in layer 4 of MJ10.

Table 2 The Color Stratigraphy, Polychromic Structures and the Pigments.

Red Pigments

Various red hues can be seen in cross-sections by OM (Fig. 1). In MJ3 the thick dark red paint layer (about 60 μm) was composed of red ochre (Fe2O3). Silicate and black soot grains were also found, which suggests that this pigment was not well purified before application. Compared with the dark red hue in MJ3, red ochre coupled with plenty of lead white (2PbCO3∙Pb(OH)2) results in the pale red hue in layer 4 of MJ4. A small amount of quartz (SiO2) and soot, which may be introduced by red ochre, were also found. The orange-red man-made pigment, minium (2PbO∙PbO2), was identified in layer 2 of MJ4 and that was why this layer showed a pale orange-red hue. Minium is unstable and can easily change into black plattnerite (PbO2), white lead white (PbCO3), lead(II) sulfate or lead (II) sulfide (Sister & Minopoulou, 2009) under the influence of some environmental factors like lightening, humidity, microbial contamination, and gas from the atmosphere (Aze et al., 2008; Daniilia et al., 2008). MJ10 is a typical sample to show instability of minium. In this sample, a thin orange-red layer of minium is covered by black plattnerite and the sample surface turned into dark red as shown in Table 1. In addition, a complete discoloration of minium was discovered in layer 3 of MJ7 where the black color could be seen. Due to the extensive usage of this pigment, especially in artworks on Silk Road, the chromatic changes (from orange-red to dark red or even black) have affected the esthetic effect of the paintings (Su et al., 2000). Another red pigment, cinnabar (HgS), was determined in MJ6. The mixture of ochre and a small amount of cinnabar produced the orange-red hue in layer 9, whereas the pure white color of lead white joined with cinnabar exhibited the scarlet color in layer 7.

Blue and Green Pigments

The cross-sectional examination in Fig. 3 is for one blue sample (MJ1) and two green samples (MJ5 and MJ7). The thin blue layer in MJ1 is composed of lazurite (Na8[Al6Si6O24]S n , n=2–3) as testified by EDS and RS (Fig. 4a) analysis. The elemental composition in the bright small particle in the BSE image (42.5 wt% Fe and 26.7 wt% S in point 2) suggests the existence of pyrite (FeS2), which is an associated mineral of lazurite in the semiprecious stone lapis lazuli. The discovery of pyrite together with the large size and angular shape of lazurite (point 1 in Fig. 3a) demonstrates the natural origin of this expensive blue pigment (Aceto et al., 2013).

Figure 3 The cross-sections of a blue sample (a) MJ1 and two green samples (b) MJ5, and (c) MJ7 by optical observation (left) and BSE observation (right), the spots where wt% are given are marked with bullets in the BSE images.

Figure 4 Raman spectra of (a) lazurite in layer 2 of MJ1, (b) atacamite in layer 7 of MJ5, (c) arsenic trioxide in layer 5 of MJ7, and (d) calcium oxalate in layer 4 of MJ7.

The green pigment used in the superficial layer of MJ5 and MJ7 is atacamite (CuCl2∙3Cu(OH)2). The transparent green color and the globular shape of atacamite in MJ6 show a synthetic origin from rustiness of copper (Cauzzi et al., 2013). As one of the earliest artificial pigments, atacamite was found in different relics such as the Kizil Grottoes (Baicheng) (Su et al., 2000), Mogao grottoes (Dunhuang), and Yungang Grottoes (Datong) (Xia, 2006). In MJ7, the minor content of As (1.4 wt%) and the typical sharp peaks by RS on the yellowish grains reveal the presence of arsenic trioxide (As2O3), which may be the transformation product of orpiment (As2S3) under the influence of lighting and moisture.

White and Black Pigments

The cross-sectional examination in Fig. 5 is for a white sample (sample MJ2) and two black samples (sample MJ8 and MJ9). In MJ2 EDS analysis of point 1 (see Supporting 3 in Supplementary Material) shows that the element ratio of K, Al, and Si is approximately 1:1:3, indicating that the white pigment is potassium feldspar (K2O∙Al2O3∙6SiO2).

Supplementary Supporting 3

Supplementary Supporting 3 can be found online. Please visit

Figure 5 The cross-sections of a white sample (a) MJ2 and two black samples (b) MJ8, and (c) MJ9 by optical observation (left) and BSE observation (right), the spots where wt% are given are marked with bullets in the BSE images.

As for sampled MJ8 and MJ9, the bright hue in BSE images and the large amount of Pb with EDS analysis suggests transformation from minium to plattnerite, which is the same as layer 3 in MJ7. However, MJ9 is covered with a fitful layer that cannot be observed by OM but shows a gray hue and distinguishes itself clearly from the beneath layer with BSE imaging. The main elements of Si and O in point 1 indicate this sample surface was polluted with sand and the additional surface image of MJ9 shown in Supporting 4 in Supplementary Material also validates this phenomenon.

Supplementary Supporting 4

Supplementary Supporting 4 can be found online. Please visit

The Priming Layer

Sample MJ1–MJ4 has no priming layer and the pigments are applied directly on the mud ground layer, whereas samples MJ5–MJ10 have priming layers with a thickness ranging from 10 to 50 μm. Silicates were found in most priming layers with the major components of Si, Al, O, and minor amounts of Na, K, Ca, Mg, P, Cl (see Supporting 1–3 in Supplementary Material). Exceptionally, chalk was discovered in layer 4 of MJ5 with the major elements of Ca, S, and O; lead white was found in the two priming layers of MJ7; zircon, which was discovered in all samples by EDS analysis, is a common accessory mineral in soil and rock. Moreover, calcium oxalate (CaC2O4) was found by typical Raman peaks (Fig. 4d) in layer 4 of MJ7 where a large amount of organic binder was also found (see Binders section). This finding certifies the statement that the oxalic acid could be related to chemical decomposition of organic materials like the binders (Prinsloo, 2007; Veneranda et al., 2014).

Supplementary Supporting 1–3

Supplementary Supporting 1–3 can be found online. Please visit


In order to avoid the color and fluorescence interference from different dyes, several pieces of cross-section of the same sample were made at the very beginning of the test. Acrylic resin, which can solidify at room temperature, was chosen as the embedding resin since heat or UV may denature the proteinaceous materials and if the configuration of the epitope of a protein is changed, the specific antigen–antibody reaction will not happen and a false-negative result may occur. Only a small amount of sample was embedded and after about 10 h, before it was cured into hard and crisp blocks, the resins were cut into several thin pieces for the following experiments so that the amount of sample is limited as much as possible. This is a micro-destructive method and can be used to deal with the investigation of invaluable works.

During IFM experiments, we found that auto-fluorescence is unavoidable (Vagnini et al., 2008; Sandu et al., 2012 b ). In order to avoid such interference, both images before and after fluorescence-labeled treatment at the same location are shown in Figures 6 and 7. Meanwhile, Alexa Fluor 488 was chosen as the only fluorochrome in IFM since its bright green fluorescence, shown in Figures 6g and 6j, could be easily observed and distinguished from the faint yellow-green auto-fluorescence shown in Figures 6f and 6i.

Figure 6 The staining test and IFM results on cross-sections of sample MJ7. Images taken under visible light, before (a) and after (b) staining with Oil Red O; images of fluorescent emission taken under blue light excitation before (c) and after (d) staining with Sypro Ruby; the IFM detection of mammalian collagen (e–g); and chicken ovalbumin (h–j), the cross-section images taken under visible light (e, h), fluorescent emission images taken under blue light excitation before (f, i) and after (g, j) IFM staining.

Figure 7 The staining test and IFM results on cross-sections of sample MJ1. Images taken under visible light, before (a) and after (b) staining with Oil Red O; images of fluorescent emission taken under blue light excitation before (c) and after (d) staining with Sypro Ruby; the IFM detection of mammalian collagen (e–g); and chicken ovalbumin (h–j), the cross-section images taken under visible light (e, h), fluorescent emission images taken under blue light excitation before (f, i) and after (g, j) IFM staining.

The results of staining tests and IFM are shown in Table 3. As classic representatives, the cross-sectional observations of sample MJ7 and MJ1 before and after tests are shown in Figures 6 and 7. In the case of MJ7, after Oil Red O staining, the upper painting and priming layer as well as the lower priming layer show red hue under visible light (Figs. 6a, 6b), which indicates the presence of oily materials. Meanwhile, by staining with Sypro Ruby, bright orange fluorescence, which signifies the presence of proteinaceous materials, can be seen in the upper priming layer, whereas relatively weak orange fluorescence can be observed in the lower priming layer (Figs. 6c, 6d). Moreover, it can be gathered from IFM results in Figures 6e to 6j that animal glue is in the upper painting layer, priming layer, and the lower priming layer; whereas egg white only exists in the upper priming layer. In the case of MJ1, oily materials were not found (Figs. 7a, 7b), but proteinaceous materials were detected in both the painting and mud layers (Figs. 7c, 7d). The results of IFM validates that animal glue was used in the mud layer (Figs. 7e–7g) and egg white was applied in the painting and mud layers (Figs. 7h–7j).

Table 3 The Staining Test and Immunofluorescence Microscopy Results of Samples from Maiji Mountain Grottoes.

The results not shown in this table are negative to all binders.

a (−) No red hue; (+) slight red hue; (++) red hue.

b (−) No fluorescence; (+) slight fluorescence; (++) fluorescence; (+++) strong fluorescence.

It can be concluded from the case of MJ7 that oily materials share the same distribution as proteinaceous materials. The same phenomenon, which was observed in the MJ8 sample (Table 3), but not in the others like MJ1 shown in Figures 7a–7d, can be explained by the reason that binders such as animal glue, milk, and egg yolk contain both proteinaceous and oily materials (animal fat). However, the dark color of the painting and mud layer or little content of oily material produced negative signals of the fat test in sample MJ1 and also other samples including MJ3, MJ5, and MJ10 and two red layers in sample MJ6.

The results of Sypro Ruby stain and IFM showed high consistency. First, the exact type of protein, whose existence had been validated by Sypro Ruby stain beforehand, was clearly determined except in sample MJ6, which may contain another type of protein, such as fish collagen and casein (milk). Second, the intensity of fluorescence, which signifies the relative amount of the analyzed materials, was highly consistent between these two analytical techniques as can be seen in Figures 6d, 6f, and 6g. All these phenomena mentioned above increased reliability of the results in this study. It can be concluded that animal glue and egg white exist in most samples and in some cases they were used together in order to achieve a better binding effect.

Painting Technique

By only viewing the cross-sections of sample MJ1, we may mistakenly think that polychrome relics in Maiji Mountain Grottoes were painted by the Fresco technique since pigment particles were inserted into the mud layer. However, the discovery of collagen and ovalbumin controverted this viewpoint. In fact, based on the finding of organic binders and the observed clear border between the priming and painting layers in most samples, it can be concluded that the painting technique is a typical Secco technique, in which pigments are mixed with organic binders and then applied on dry lime-based plaster or a smooth mud layer. Such a situation was also encountered when investigating the mural painting in cave 3 of Mogao Grottoes in Dunhuang (Guo, 2004), owing to the disappearance of the lime-based priming layer, and meanwhile the large size and amount of sand that made the surface course and induced the pigment particles to penetrate into a mud layer easily. Actually, fresco painting has not yet been found in polychrome relics in China.


In this study, ten polychromic samples from Maiji Mountain Grottoes have been comprehensively investigated by the integrated analytical techniques of OM, SEM-BSE, SEM-EDS, μ-RS, staining test, and IFM. Various red colors have been found and the pigments are red ochre (Fe2O3), minium (2PbO∙PbO2), and cinnabar (HgS) used individually or in mixture. Lazurite (Na8[Al6Si6O24]Sn) is the main component of the utilized blue pigment. The green pigment is atacamite (CuCl2∙3Cu(OH)2). It is proposed that arsenic-containing compounds like As2S3 was used together with copper pigments for adjusting the visual sense. The black surface is plattnerite which was discolored from minium. The white surface was painted with potassium feldspar (K2O∙Al2O3∙6SiO2). The polychrome relics studied in this work had been repainted several times over centuries, which is the reason for the complex polychrome with multiple layers (priming, painting, and repainting layers) on the relics. Binders in the polychrome paints were successfully identified through a series of experiments including oily materials identified by Oil Red O stain, proteinaceous materials revealed by Sypro Ruby stain, and protein determined by IFM. The distribution of stain color or fluorescence revealed the exact location of binders, and the intensities of fluorescence indicate the relative amounts of binders in multilayer structures of polychrome relics. The original materials of proteinaceous binders were determined as animal glue and egg white, which were distributed in most painting and priming layers of the polychromic samples. The combination of the employed techniques is effective for binder identification in artworks, especially in multilayer artworks. Finally, we draw the conclusion that the painting technique used in Maiji Mountain Grottoes is the Secco technique.


The authors are grateful to Professor Guofeng Wei (Department of History, Anhui University) for help with Raman spectrum analysis. The National Basic Research Program of China (Grant Number 2012CB720902), National Innovation League (Zhejiang Province) in Cultural Heritage Conservation Science and Technology (Grant Number [2012] 878) are greatly acknowledged for their financial support.


Aceto, M., Agostino, A., Fenoglio, G. & Picollo, M. (2013). Non-invasive differentiation between natural and synthetic ultramarine blue pigments by means of 250-900nm FORS analysis. Anal Methods 5(16), 41844189.
Aze, S., Vallet, J.M., Detalle, V., Grauby, O. & Baronnet, A. (2008). Chromatic alterations of red lead pigments in artworks: a review. Phase Transit 81(2–3), 145154.
Bardelli, F., Barone, G., Crupi, V., Longo, F., Majolino, D., Mazzoleni, P. & Venuti, V. (2011). Combined non-destructive XRF and SR-XAS study of archaeological artefacts. Anal Bioanal Chem 399(9), 31473153.
Bonaduce, I., Blaensdorf, C., Dietemann, P. & Colombini, M.P. (2008). The binding media of the polychromy of Qin Shihuang’s Terracotta Army. J Cult Herit 9(1), 103108.
Burgio, L. & Clark, R.J.H. (2001). Library of FT-Raman spectra of pigments, minerals, pigment media and varnishes, and supplement to existing library of Raman spectra of pigments with visible excitation. Spectrochim Acta A Mol Biomol Spectrosc 57(7), 14911521.
Cappitelli, F. & Koussiaki, F. (2006). THM-GCMS and FTIR for the investigation of paints in Picasso’s Still Life, Weeping Woman and Nude Woman in a Red Armchair from the Tate Collection, London. J Anal Appl Pyrol 75(2), 200204.
Cartechini, L., Vagnini, M., Palmieri, M., Pitzurra, L., Mello, T., Mazurek, J. & Chiari, G. (2010). Immunodetection of Proteins in Ancient Paint Media. Acc Chem Res 43(6), 867876.
Cauzzi, D., Chiavari, G., Montalbani, S., Melucci, D., Cam, D. & Ling, H. (2013). ). Spectroscopic and chromatographic studies of sculptural polychromy in the Zhongshan Grottoes (R.P.C.). J Cult Herit 14(1), 7075.
Daniilia, S., Minopoulou, E., Demosthenous, F.D. & Karagiannis, G. (2008). A comparative study of wall paintings at the Cypriot monastery of Christ Antiphonitis: one artist or two? J Archaeol Sci 35(6), 16951707.
Dong, Y. (1983). The stages of Maiji Mountain Grottoes. Cult Relics 29(6), 1830. 103–104.
Franquelo, M.L., Duran, A., Castaing, J., Arquillo, D. & Perez-Rodriguez, J.L. (2012). XRF, mu-XRD and mu-spectroscopic techniques for revealing the composition and structure of paint layers on polychrome sculptures after multiple restorations. Talanta 89, 462469.
Guo, H. (2004). The identification of ancient Secco and Buon Frescoes. Cult Relics Central China 38(2), 7680.
Hu, W., Zhang, K., Zhang, H., Zhang, B. & Rong, B. (2015). Analysis of polychromy binder on Qin Shihuang’s Terracotta Warriors by immunofluorescence microscopy. J Cult Herit 16(2), 244248.
Huang, W. (1989). The history of Maiji Mountain Grottoes. Cult Relics 35(3), 8389. 96, 104-105.
Kuckova, S., Anca Sandu, I.C., Crhova, M., Hynek, R., Fogas, I. & Schafer, S. (2013). Protein identification and localization using mass spectrometry and staining tests in cross-sections of polychrome samples. J Cult Herit 14(1), 3137.
Mancini, D., Tournie, A., Caggiani, M.-C. & Colomban, P. (2012). Testing of Raman spectroscopy as a non-invasive tool for the investigation of glass-protected miniature portraits. J Raman Spectrosc 43(2), 294302.
Mazzocchin, G.A., Agnoli, F., Mazzocchin, S. & Colpo, I. (2003). Analysis of pigments from Roman wall paintings found in Vicenza. Talanta 61(4), 565572.
Navas, N., Romero-Pastor, J., Manzanoa, E. & Cardell, C. (2008). Benefits of applying combined diffuse reflectance FTIR spectroscopy and principal component analysis for the study of blue tempera historical painting. Anal Chim Acta 630(2), 141149.
Palmieri, M., Vagnini, M., Pitzurra, L., Rocchi, P., Brunetti, B.G., Sgamellotti, A. & Cartechini, L. (2011). Development of an analytical protocol for a fast, sensitive and specific protein recognition in paintings by enzyme-linked immunosorbent assay (ELISA). Anal Bioanal Chem 399(9), 30113023.
Piovesan, R., Siddall, R., Mazzoli, C. & Nodari, L. (2011). The Temple of Venus (Pompeii): a study of the pigments and painting techniques. J Archaeol Sci 38(10), 26332643.
Prinsloo, L.C. (2007). Rock hyraces: a cause of San rock art deterioration? J Raman Spectrosc 38(5), 496503.
Rosi, F., Burnstock, A., Van den Berg, K.J., Miliani, C., Brunetti, B.G. & Sgamellotti, A. (2009). A non-invasive XRF study supported by multivariate statistical analysis and reflectance FTIR to assess the composition of modern painting materials. Spectrochim Acta A-Mol Biomol Spectrosc 71(5), 16551662.
Sandu, I.C.A., Roque, A.C.A., Matteini, P., Schaefer, S., Agati, G., Correia, C.R. & Fernandes Pacheco Viana, J.F. (2012a). Fluorescence recognition of proteinaceous binders in works of art by a novel integrated system of investigation. Microsc Res Techniq 75(3), 316324.
Sandu, I.C.A., Schaefer, S., Magrini, D., Bracci, S. & Roque, C.A. (2012b). Cross-section and staining-based techniques for investigating organic materials in painted and polychrome works of art: A review. Microsc Microanal 18(4), 860875.
Sister, D. & Minopoulou, E. (2009). A study of smalt and red lead discolouration in Antiphonitis wall paintings in Cyprus. Appl Phys A Mater Sci Process 96(3), 701711.
Su, B., Li, Z., Ma, Z., Li, S. & Ma, Q. (2000). A study of pigments in murals of Kizil Grottoes. Dunhuang Res 63(1), 6575.
Vagnini, M., Pitzurra, L., Cartechini, L., Miliani, C., Brunetti, B.G. & Sgamellotti, A. (2008). Identification of proteins in painting cross-sections by immunofluorescence microscopy. Anal Bioanal Chem 392(1–2), 5764.
Veneranda, M., Irazola, M., Pitarch, A., Olivares, M., Iturregui, A., Castro, K. & Manuel Madariaga, J. (2014). In-situ and laboratory Raman analysis in the field of cultural heritage: The case of a mural painting. J Raman Spectrosc 45(3), 228237.
Wang, N., He, L., Egel, E., Simon, S. & Rong, B. (2014). Complementary analytical methods in identifying gilding and painting techniques of ancient clay-based polychromic sculptures. Microchem J 114, 125140.
Wei, H. (2011). A literature review on the study on Maiji Mountain Grottoes in Tianshui. Lanzhou: Northwest Normal University.
Wei, S., Ma, Q. & Schreiner, M. (2012). Scientific investigation of the paint and adhesive materials used in the Western Han dynasty polychromy terracotta army, Qingzhou, China. J Archaeol Sci 39(5), 16281633.
Xia, Y. (2006). Polarized light microscopy, application on ancient Chinese pigments’ study and related database establishing. Xi’an: Northweatern University.
Zhou, G. (1991). The analysis of inorganic pigments from the mural paintings and polychromic sculptures of Maiji Mountain Grottoes by X-ray Diffractions. Archaeology 35(8), 744755.